Depress ion and Psychologica l Dis tress in Pat ients During theYear Af ter Curat i ve Resect ion of NonSmal l -Cel l Lung Cancer

By Yosuke Uchitomi, Ichiro Mikami, Kanji Nagai, Yutaka Nishiwaki, Tatsuo Akechi, and Hitoshi Okamura

Purpose: There have been few psychosocial studies ofpatients after curative resection of nonsmall-cell lung can-cer (NSCLC). The purpose of this study was to clarify theclinical course of depression and psychological distress ofsuch patients during the year after surgery and to identifypredictors of their long-term outcome.

Patients and Methods: A total of 212 patients completedassessments during a 12-month follow-up period after cura-tive resection of NSCLC. Psychological measurements at 1, 3,and 12 months after surgery were conducted using the Struc-tured Clinical Interview for Diagnostic and Statistical Manualof Mental Disorders (ed 3), Revised, and the Profiles of MoodStates (POMS) scale. Univariate and multivariate analyseswere used to identify predictors of psychological outcomeaccording to these two methods of assessment.

Results: The prevalence of depression did not changeduring the year after curative resection (range, 4.7% to8.0%). The total POMS score was also unaltered during the

year after surgery: the anger-hostility (P < .001) and ten-sion-anxiety subscale scores (P < .026) had increased at 12months, but the vigor-activity subscale score had also in-creased (P < .001). All predictors of psychological outcomeat 12 months included a depression episode after the diag-nosis of lung cancer or at 1 month after surgery. Less-educated status was also a significant predictor of depres-sion at 12 months.

Conclusion: These results suggest the need for psychos-ocial support even after curative resection of NSCLC andindicate that an approach that includes repetitive perioper-ative assessment of depression and careful attention toless-educated patients might be of benefit to patients inameliorating depression and psychological distress duringthe year after curative resection.

J Clin Oncol 21:69-77. 2003 by AmericanSociety of Clinical Oncology.

L UNG CANCER is the most common form of cancer and themost common cause of cancer-related deaths in theworld.1,2 Since the 1980s, extensive quality-of-life (QOL) re-search has been introduced in clinical trials on lung cancerpatients,3,4 but few of these studies have examined psychologicaldistress after curative resection of nonsmall-cell lung cancer(NSCLC).5-7 Although surgical treatment for early-stageNSCLC is generally considered curative, the outcome of NSCLCis still unsatisfactory: The postoperative 5-year survival rate is79.0% for patients with pathologic stage IA disease, 59.7% forthose with stage IB disease, 56.9% for those with stage IIAdisease, and 45.0% for those with stage IIB disease.8 There aresurvivors after curative resection of NSCLC, despite the overallgrim statistics. The psychological distress of NSCLC patientsafter surgery also seems to be less well understood in the contextof curative cancer treatment.3,4

Psychological distress, including depression, is an essentialelement of the QOL of cancer patients and, thus, depression hasa great impact on their QOL.9-11 Depression may be associatedwith treatment decision making, such as choosing chemothera-peutic agents,12 and with the shorter survival of lung cancerpatients.13,14 Nevertheless, other studies have revealed thatmedical staff are poor at detecting emotionally distressed ordepressed cancer patients.15-17

Depression is common in cancer patients and occurs through-out the course of their illness.18-20 Previous reports21-25 ondepression after a diagnosis of lung cancer have revealed that15% to 44% of patients experience some form of depression,including major depression and adjustment disorders with de-pressive mood, on the basis of Diagnostic and Statistical Manualfor Mental Disorders, Revised (DSM-III-R) criteria,26 or clini-cally probable and borderline depression, on the basis of theHospital Anxiety and Depression Scale (HADS).27 Because lungcancer and its treatment varies with histologic type and disease

stage, previous studies have merely shown that some patientsmay experience depression after curative resection of NSCLC.We therefore first examined the prevalence of depression usingthe Structured Clinical Interview for DSM-III-R (SCID)28 in alarge, homogeneous, consecutive, prospectively designed sam-ple of NSCLC patients during the 3 months after curativeresection, and found that the 1-month prevalence of depression at1, 2, and 3 months after surgery was 9.0%, 9.4%, and 5.8%,respectively.29 Although the prevalence of depression was lowerthan in lung cancer patients as a whole, including patients withadvanced and small-cell lung cancer, it was not so low that it wasnegligible (0.9% to 3.7% in the general populations).30-32

Factors associated with psychological distress, including de-pression, in various cancers have included physical variables,such as pain, fatigue, other symptom burden, and poor performancestatus (PS),25,33-35whereas other studies have indicated sociodemo-graphic and psychosocial variables, such as younger age36 andsocial support, including marital status.37 Because physical vari-ables, such as PS and dyspnea, would be expected to improve by 6or 9 months after surgery,5-7 psychosocial factors may more

From the Psycho-Oncology Division, National Cancer Center ResearchInstitute East, and Psychiatry Division, National Cancer Center HospitalEast, Kashiwa, Chiba; Psychiatry Division, National Shikoku Cancer Cen-ter, Matsuyama, Ehime; Thoracic Oncology Division, National CancerCenter Hospital East, Kashiwa, Chiba; and Division of OccupationalTherapy, Institute of Health Sciences, Hiroshima University School ofMedicine, Hiroshima, Japan.

Submitted December 27, 2001; accepted August 20, 2002.Address reprint requests to Yosuke Uchitomi, MD, PhD, Psycho-Oncology

Division, National Cancer Center Research Institute East, Kashiwanoha6-5-1, Kashiwa, Chiba 277-8577, Japan; email: yuchitom@east.ncc.go.jp.

2003 by American Society of Clinical Oncology.0732-183X/03/2101-69/$20.00

69Journal of Clinical Oncology, Vol 21, No 1 (January 1), 2003: pp 69-77DOI: 10.1200/JCO.2003.12.139

strongly predict depression and psychological distress after curativeresection in NSCLC patients than physical factors.

Previous studies that examined the prevalence of depressionand factors associated with depression in lung cancer patientshave entailed certain drawbacks, such as small sample size,21-

23,25 a heterogeneous subject population that included subjectswith all histologic types and all disease stages,21-23 measurementof depression by self-report measures with limited accura-cy,23,24,34 and design as a cross-sectional or short-term follow-upstudy.21-25,29,34 In this study, we chose to use a structured clinicalinterview, after which clinical intervention for psychiatric disor-ders had to be recommended by feedback to the attendingphysician as both a means of clarifying the clinical course ofdepression by rigorous methods during the year after curativeresection of NSCLC and a means of identifying factors predic-tive of long-term outcome. We also used reliable, valid, self-administered instruments to assess psychological distress duringthe year after surgery.

PATIENTS AND METHODS

Participants

Consecutive newly diagnosed patients were invited to participate in thestudy after curative resection of NSCLC conducted at the Thoracic OncologyDivision, National Cancer Center Hospital East, Kashiwa, Japan.

The eligibility criteria were 18 years of age or older; awareness of thediagnosis of cancer; ability to speak Japanese; standard surgical treatment(lobectomy or pneumonectomy with mediastinal lymph node dissection); noevidence of brain tumor on computerized tomography or magnetic resonanceimages of the head; no history of or current chemotherapy, immunotherapy, orradiation therapy; no active concomitant cancer; curative resection on the basisof pathology reports of an International Union Against Cancer (UICC) diseasestage of pT1 to pT3, pN0/1, pM0;38 and no severe medical conditions at 1 monthafter surgery.

Procedure

The study was approved by the institutional review board of the NationalCancer Center, Japan. Each patient was fully informed of the purpose of thestudy before written consent was obtained. Some demographic and clinicalcharacteristics, including age, sex, type of surgery (lobectomy or pneumo-nectomy), preoperative percent vital capacity (%VC), preoperative percentforced expiratory volume in 1 second (%FEV1), and preoperative PS, wereobtained from the patients medical charts, whereas education level, maritalstatus, information about whether the patient lived alone, employment status,smoking status, and history of nicotine dependence were obtained duringsemistructured interviews conducted in the interview room on the ward bytwo psychiatrists (Y.U., I.M.) before discharge (median interval aftersurgery, 13 days). Preoperative smoking status was based on patientsself-reports. Current smokers were defined as those who smoked at thetime of surgery or had quit smoking within the previous year, andex-smokers were defined as those who had quit smoking 1 or more yearsbefore surgery. History of nicotine dependence was assessed according tothe DSM-III-R.26

Pathologic disease stage was obtained from the patients medical charts,and PS, pain, and dyspnea were based on semistructured interviews con-ducted in the interview room of the Outpatient Service, Thoracic OncologyDivision, National Cancer Center Hospital East, by two psychiatrists (Y.U.,I.M.) 1 month after surgery. The evaluation of PS was based on the EasternCooperative Oncology Group criteria. Pain and dyspnea were graded on afour-point verbal scale: (0) absent, (1) mild, (2) moderate, and (3) severe.

At 1, 3, and 12 months after surgery (median intervals after surgery, 31days, 92 days, and 377 days, respectively), a semistructured interview wasconducted to obtain information on demographic and clinical characteristicsand was immediately followed by psychological measurements using theSCID28 and the Profiles of Mood States (POMS) scale.39 A blood specimenwas collected at 12 months after surgery to determine smoking statuschemically.

Most measures at baseline, including measures of depression and thePOMS, could not be performed before surgery for NSCLC, mainly because wejudged that the baseline interview before surgery would be stressful to mostpatients. They had only been informed by the hospital of the day they would beadmitted for surgical treatment of NSCLC 1 or 2 days in advance. Anotherreason was the practical problem of not being able to interview patients becauseof the brief interval between the time they were admitted and the operation(usually 2 or 3 days). We therefore scheduled the baseline interview at 1 monthafter surgery, by which time patients had been informed of the final pathologicdiagnosis.

When patients were diagnosed with depression, we recommended psychi-atric consultation to the attending physician. When patients agreed, routinepsychiatric consultations were provided. Our study was, therefore, limitedbecause we were unable to observe the natural course of depression andpsychological distress during the year after curative resection.

Measures

At 1, 3, and 12 months after surgery, a psychiatrist (I.M.) used the SCIDduring an interview to evaluate the patients for depression during theprevious month. Patients were also evaluated using the SCID for history ofdepression. Whether depression was present before the patient was told ofthe diagnosis of NSCLC and between when the patient was told of thediagnosis and surgery was determined by patient report. The number ofmajor depressive episode items according to the DSM-III-R was obtained bysumming the number of the following nine items that were present at thetime: (1) depressive mood, (2) diminished interest or pleasure, (3) change inbody weight or appetite, (4) insomnia or hypersomnia, (5) psychomotoragitation or retardation, (6) fatigue or loss of energy, (7) feelings ofworthlessness or guilt, (8) loss of concentration or indecisiveness, and (9)suicidal thoughts. The diagnosis of depression was made according to thecriteria for major and/or minor depression. Major depression was diagnosedwhen five or more items, including either depressive mood or diminishedinterest or pleasure, were present; minor depression was diagnosed when twoto four items, including depressive mood or diminished interest or pleasure,were present.

As expected, it was difficult to determine whether somatic symptoms, suchas appetite loss, insomnia, and fatigue, were attributable to depression orsurgical treatment. There are four approaches to diagnose depression of themedically ill patient: to exclude somatic symptoms (exclusive); to substitutepsychological symptoms for somatic symptoms (substitutive); to countsomatic symptoms toward a depression diagnosis unless symptom is clearlyand fully accounted for by a general medical condition (etiologic); and toinclude somatic symptoms (inclusive).40 Because the rater must make ajudgment about the cause of individual symptoms, the etiologic approachmay be less reliable than the exclusive, substitutive, and inclusive approach.We elected to use an inclusive diagnostic approach (to prevent underdiag-nosis of depression), which is considered to be the most important approachin the clinical setting, although this approach may result in overestimation ofdepression in the medically ill patient. The reliability of the interview ratingswas determined by having a second rater (Y.U.) attend a random sampleconsisting of 30 interviews (4.8%). The interrater agreement (kappa) valuesfor the diagnosis of depression, major depression, and minor depression were0.78, 1.00, and 0.65, respectively. Ratings for each of the nine individualitems were also reliable, with the kappa values ranging from 0.65 to 1.00.

Patients psychological distress was assessed using the POMS scale, a65-item self-rating scale measuring six emotional states (tension-anxiety,depression-dejection, anger-hostility, vigor-activity, fatigue-inertia, and con-fusion-bewilderment), and its summary score, the total mood disturbance(TMD) score.39 The validity and reliability of the Japanese version of thePOMS have been confirmed.41

Plasma cotinine concentrations were blindly determined in duplicate by ahigh-performance liquid chromatography method with ultraviolet detectionaccording to the method described by Hariharan et al.42 The minimumquantifiable concentration was 2 ng/mL. Patients with a cotinine concentra-tion of less than 20 ng/mL were classified as nonsmokers, and patients withconcentration of 20 ng/mL were classified as smokers.

Statistical Analysis

Standard descriptive statistics were used to characterize the distributionsof the diagnosis of depression and the POMS scores at the 1-, 3-, and12-month follow-up visits. Intergroup comparisons of categorical and non-

70 UCHITOMI ET AL

parametric variables were performed using the 2 test and the Wilcoxon ranksum test, respectively. Associations between continuous variables wereexamined by calculating Spearman rank correlation coefficients. Thechanges over time in the POMS scores were tested using the Wilcoxon testor the Friedman test. The percentages of patients with depression werecompared across the two or three time points by the McNemar test or theCochran Q test.

The predictors examined were age, sex, education, marital status, whetherthe patient lived alone, employment status, type of surgery, preoperative%VC, preoperative %FEV1, preoperative PS, preoperative smoking status,smoking status at 12 months after surgery, history of nicotine dependence,history of depression before being told of the diagnosis of NSCLC, historyof depression between the diagnosis of NSCLC and surgery, and PS, pain,dyspnea, depression, and POMS-TMD score at 1 month after surgery.

After the results for the correlations between biomedical and psychosocialvariables related to depression and the POMS-TMD score at 12 months aftersurgery in the univariate comparisons (P .25) had been examinedcarefully, they were entered into multivariate models to examine thepredictive factors of depression and the POMS-TMD score at 12 monthsafter surgery. A logistic regression analysis with backward elimination wasthen used with alpha 0.05 as the significance criterion to select the finalmodels of depression, and multiple regression analysis with backwardelimination was performed to examine the predictors of the POMS-TMDscore at 12 months. In all statistical evaluations, P values of 0.05 or less wereconsidered indicative of significant differences. SPSS 10.0J for Windowsstatistical software (SPSS Japan Institute Inc., Tokyo, Japan, 2000) was usedfor all data analyses.

RESULTS

Curative resection was confirmed by the final pathologicreport of disease stage in 262 (86.5%) of the 303 NSCLCpatients who underwent standard surgery between June 1996 andApril 1999. At the time of the 1-month interview, three patientscould not be contacted and 20 refused to participate in the study(10 of them because of the psychological burden, three becauseof the psychological and physical burden, four because of thephysical burden, and three for unknown reasons). By the time ofthe 3-month follow-up, two patients had been lost to follow-up,two had died, and seven refused to participate (four because ofthe psychological burden, one because of the physical burden,one because of time constraints, and one for unknown reasons).By the time of the 12-month follow-up, five patients had beenlost to follow-up, seven had died, three were too ill to participate,and one refused to participate for unknown reasons. Thus, thefinal study group consisted of 212 patients, representing 80.9%of all eligible patients.

Data for the eligible patients who had undergone successfulsurgical resection but did not participate in the study (n 50,19.1%) were available in regard to age, sex, education, maritalstatus, preoperative PS, and clinical disease stage. More of thenonparticipants were unmarried (26.1% of nonparticipants v16.0% of participants, P .018), and significantly more non-participants had advanced clinical stage disease (stage IIIA:30.0% of nonparticipants v 6.6% of participants, P 0.001), butthe two groups did not differ with respect to any other availabledata. Of the 17 nonparticipants who refused to participate at thetime of the follow-up examinations for reasons that includedpsychological burden, one was diagnosed with major depressionafter referral to the Psychiatry Division. After psychiatric referral,one of the 10 patients who could not be contacted was diagnosedwith major depression, and one of the three who were judged to betoo ill to participate was diagnosed with minor depression.

The mean age SD of the 212 cancer patients who partici-pated in the study was 62.1 10.8 years (median, 63.5; range,

22 to 83), 39.6% were female, 84% were married, 7% livedalone, and 33% had a junior high school education or less.Smoking status was as follows: current smokers, 80 (37.7%);ex-smokers, 54 (25.5%); and nonsmokers, 78 (36.8%). Of thecurrent and ex-smokers, 89 participants had a history of nicotinedependence. Preoperative PS was grade 0 in 148 patients(69.8%), grade 1 in 63 patients (29.7%), and grade 2 in onepatient (0.5%). Pathologic stage I disease was the most commonstage (n 165; 77.8%), followed by pathologic stage II (n 33,15.6%), and pathologic stage IIIA (n 14; 6.6%). Lobectomywas performed in 203 patients (95.8%) and pneumonectomy innine (4.3%). Preoperative respiratory function was generallygood: 19 patients (9.0%) had a %VC below 80%, and 44 (20.8%)had a %FEV1 below 70%. According to the results of the serumcotinine assay 1 year after surgery, 21 of the current smokers(n 80) and none of the ex-smokers (n 54) had continuedto smoke.

Table 1 shows preoperative and 12-month postoperativerespiratory function in the form of %VC and %FEV1; preoper-ative, 1-, 3-, and 12-month scores on the PS scales; andprevalence of depression and POMS scores at 1, 3, and 12months postoperatively. The results show a statistically signifi-cant reduction in %VC (15%) at 1 year after surgery, but nochange in %FEV1. There were significant reductions in PSscores after surgery; PS at 12 months returned to the preopera-tive level. Pain and dyspnea significantly decreased at 1, 3, and12 months after surgery (pain, 2.50 0.66, 2.16 0.64, and1.83 0.68, respectively, P .001; dyspnea, 2.30 0.68,2.10 0.69, and 1.81 0.68, respectively, P .001).

The 1-month prevalence of depression at 1, 3, and 12 monthswas 8.0%, 5.2%, and 4.7%, respectively, and none of thedifferences in 1-month prevalence of depression during the yearwere significant (Table 1). The 1-month prevalence of minordepression at 1, 3, and 12 months was 3.3%, 2.4%, and 3.3%,respectively, with no significant changes during the year. The1-month prevalence of major depression at 1, 3, and 12 monthswas 4.7%, 2.8%, and 1.4%, respectively, showing a tendency todecrease, but none of the differences in 1-month prevalence ofdepression during the year were significant.

The clinical course of the depression is shown in Fig 1. Onlythree (17.6%) of the subjects with depression at 1 month (n 17) were diagnosed with depression at 12 months. Of thesubjects found not to be depressed at 1 month (n 195), sevenwere diagnosed with depression at 12 months. The number ofpatients with a history of depression before the diagnosis of lungcancer was 30 (14.2%). The number of patients with a history ofdepression after the diagnosis of lung cancer but before surgerywas 24 (11.3%).

Four of the 10 patients with major depression at 1 month andnone of the seven subjects with minor depression at 1 monthwere referred to the Psychiatry Division by the attendingphysician immediately after the 1-month interview. After psy-chiatric intervention, two of the four patients with major depres-sion at 1 month were not diagnosed with depression at 3 months,but the other two were still diagnosed with major depression andunder psychiatric treatment. At the 12-month follow-up, onepatient was still diagnosed with major depression, but the otherthree were not diagnosed with depression. Only two of the fivepatients with minor depression at 3 months were referred to thePsychiatry Division, and neither patient was diagnosed with

71DEPRESSION AFTER LUNG CANCER SURGERY

depression at the 12-month follow-up. Of the six patientsreferred to the Psychiatry Division, all received supportivepsychotherapy and/or antidepressant treatment, and five were notdiagnosed with depression at 12 months after surgery.

There was no change in the POMS-TMD score during the yearafter surgery (Table 1). Of the six subscales of the POMS,although the anger-hostility score had increased and the tension-anxiety score had increased slightly but significantly, the vigor-activity scale had also increased significantly. After the sixpatients referred to the Psychiatry Division were excluded, thePOMS-TMD score did not change during the year after surgery(POMS-TMD scores at 1, 3, and 12 months were 19.9 22.8,17.9 22.8, and 18.5 23.9, respectively, P .281, n 202).After the 22 patients given recommendations for psychiatricconsultation were excluded, the POMS-TMD score did not

change during the year after surgery (POMS-TMD scores at 1, 3,and 12 months were 17.6 20.6, 15.2 19.8, and 16.8 22.7,respectively, P .214, n 187).

The results of the univariate analysis are shown in Table 2,and age, sex, education, marital status, preoperative PS, preop-erative smoking status, history of depression before lung cancer,depression at 1 month, and POMS-TMD at 1 month were enteredinto the depression model. In addition, age, sex, type of surgery,preoperative %FEV1, history of depression before lung cancer,history of depression between the diagnosis of lung cancer andsurgery, dyspnea at 1 month, depression at 1 month, and POMS-TMD at 1 month were entered into the POMS-TMD model.

The results of the multivariate analysis are shown in Table 3.Depression and the POMS-TMD score at 1 month were signif-icant predictors of the outcome at 12 months. History of

Fig 1. Clinical course of major and mi-nor depression in nonsmall-cell lung cancerpatients during the 12 months after curativeresection (n 212).

Table 1. Respiratory Function, Performance Status, Depression, and Psychological Distress in NonSmall-Cell Lung Cancer Patients After CurativeResection (n 212)

VariablePreoperativeMean SD

12-Month*Mean SD P

%VC 104.5 16.6 89.1 15.7 .001%FEV1 75.2 9.4 75.0 11.6 .124

Preoperative Mean SD

1-Month Mean SD

3-Month Mean SD

12-Month Mean SD P P

Performance status (ECOG) 0.31 0.47 1.01 0.32 0.66 0.50 0.31 0.50 .001 .001POMS-TMD 20.8 23.3 18.2 23.1 18.7 24.2 .221 .177Tension-anxiety 7.0 4.7 7.5 4.4 7.5 4.2 .028 .063Depression-dejection 6.3 6.8 6.2 7.1 6.6 6.7 .311 .435Anger-hostility 4.3 5.4 4.8 4.8 6.0 6.0 .001 .001Vigor-activity 9.3 5.5 12.0 6.2 13.3 5.8 .001 .001Fatigue-inertia 5.9 4.5 5.6 4.5 5.4 4.7 .227 .121Confusion-bewilderment 6.6 3.5 6.1 3.5 6.4 3.7 .110 .611

1 Monthn (%)

3 Monthn (%)

12 Monthn (%) P P

Depression 17 (8.0) 11 (5.2) 10 (4.7) .203 .189Major depression 10 (4.7) 6 (2.8) 3 (1.4) .071 .065Minor depression 7 (3.3) 5 (2.4) 7 (3.3) .801 .99

*For respiratory function at 12 months, n 190.Across all months.1-month versus 12-month comparison only.For POMS at 1 and 3 months, n 210.Abbreviations: ECOG, Eastern Cooperative Oncology Group; POMS-TMD, Profiles of Mood States-total mood disturbance.

72 UCHITOMI ET AL

Table 2. Demographic and Biomedical Characteristics, and Depression and Psychological Distress in NonSmall-Cell Lung Cancer Patients at 12Months After Curative Resection: Univariate Analyses (n 212)

Characteristics n

Depression POMS-TMD Score

n 10 (%) P Mean SD P

Age (years) 65 120 8 (6.7) .192 22.1 26.8 .057 65 92 2 (2.2) 14.2 19.4

SexMale 128 4 (3.1) .199 16.4 25.9 .010Female 84 6 (7.1) 22.1 20.9

Education (years) 9 71 8 (11.3) .003 22.3 28.6 .499 9 141 2 (1.4) 16.8 21.5

MarriedNo 34 3 (8.8) .204 22.9 24.5 .252Yes 178 7 (3.9) 17.8 24.1

Living aloneYes 15 1 (6.7) .528 22.0 28.5 .770No 197 9 (4.6) 18.4 23.9

EmploymentNo 114 6 (5.3) .755 19.0 23.6 .588Yes 98 4 (4.1) 18.2 25.0

Type of surgeryLobectomy 203 10 (4.9) .99 19.3 24.4 .042Pneumonectomy 9 0 (0) 4.4 11.7

Preoperative %VC 80 19 0 (0) .605 17.1 18.1 .938 80 193 10 (5.2) 18.8 24.7

Preoperative %FEV1 70 44 2 (4.5) .99 14.2 18.9 .186 70 168 8 (4.8) 19.8 25.3

Pathologic disease stage r 0.025 .718I 165 7 (4.2) .533 18.4 24.5II 33 2 (6.1) 18.1 22.4IIIA 14 1 (7.1) 22.5 25.2

Preoperative performance status0 148 4 (2.7) .070 17.3 21.7 .5991/2 64 6 (9.4) 21.8 29.1

Performance status at 1 month0 8 1 (12.5) .325 8.8 22.9 .3211/2 204 9 (4.4) 19.0 24.2

Pain at 1 monthNone to mild 98 3 (3.1) .346 18.2 24.7 .745Moderate to severe 114 7 (6.1) 19.0 23.8

Dyspnea at 1 monthNone to mild 124 5 (4.0) .744 17.6 25.2 .216Moderate to severe 88 5 (5.7) 20.2 22.6

Preoperative smoking statusCurrent smoker 80 6 (7.5) .183 21.8 29.1 .477Ex- and nonsmoker 132 4 (3.0) 16.7 20.5

Smoking status at 12 monthsContinued smoking 21 2 (9.5) .259 24.3 30.0 .522Others 191 8 (4.2) 18.0 23.5

Smoking status at 12 months 2 (9.5) 24.3 30.0Quit smoking 59 4 (6.8) .470 20.9 29.1 .733Others 153 6 (3.9) 17.8 22.0

History of nicotine dependenceNo 123 4 (3.3) .327 16.3 20.0 .427Yes 89 6 (6.7) 21.9 28.8

History of depression before lung cancerNo 182 6 (3.3) .037 17.1 24.0 .007Yes 30 4 (13.3) 28.1 23.1

History of depression between thediagnosis of lung cancer and surgery

No 188 8 (4.3) .315 16.8 22.8 .009Yes 24 2 (8.3) 33.2 30.0

Depression at 1 monthNo 195 7 (3.6) .036 16.7 22.5 .001Yes 17 3 (17.6) 40.6 31.7

POMS-TMD score at 1 month* 28.6 19.4 .133 r .592 .00120.4 23.5

*For POMS at 1 month, n 210.The mean value SD of depression yes (n 10).The mean value SD of depression no (n 200).

73DEPRESSION AFTER LUNG CANCER SURGERY

depression between the diagnosis of lung cancer and surgery wasalso a significant predictor in the POMS model. Junior highschool education or less was a significant predictor in thedepression model.

After the six patients referred to the Psychiatry Division wereexcluded, the same predictors continued to be significant in boththe depression model (depression at 1 month, odds ratio [OR] 8.14, P .030; 95% CI, 1.23 to 53.93, junior high schooleducation or less, OR 9.18, P .009, n 206, 95% CI, 1.73to 48.74) and the POMS model (POMS-TMD at 1 month,standardized coefficient 0.47, P .001; history of depressionafter lung cancer before surgery, standardized coefficient 0.18,P .003, n 204). After the 22 patients given recommenda-tions for psychiatric consultation were excluded, junior highschool education or less did not remain a significant predictor inthe depression model (OR 5.13, , P .055, n 190, 95% CI,0.97 to 27.19) but the POMS-TMD score at 1 month continuedto be a significant predictor in the POMS model (standardizedcoefficient 0.53, P .001, n 188).

Because the predictors in both models included perioperativedepression, that is, a depression episode between the diagnosis oflung cancer and surgery or at 1 month, the data were separatedaccording to whether the patients had experienced perioperativedepression and then reanalyzed (Table 4). Of the 37 patients withperioperative depression (n 37), four had a depression episodeboth times, that is, a depression episode between the diagnosis oflung cancer and surgery and an episode 1 month after curativeresection. One of the predictors in the POMS model, the POMS

score at 1 month, continued to predict in the model of subjectswithout perioperative depression (Table 4A), but no significantpredictors remained in the depression model of subjects withoutperioperative depression (junior high school education or less,OR 4.87, P .063, n 175, 95% CI, 0.92 to 25.88). Afterthe 17 patients diagnosed with depression at 1 month wereexcluded from the depression model, junior high school educa-tion or less showed a tendency to be a predictor (OR 5.21, P .053, n 195, 95% CI, 0.98 to 27.59).

In the models of the subjects with perioperative depression(Table 4B), junior high school education or less became asignificant predictor in the POMS-TMD model. No significantpredictors remained in the depression models of the subjectswith perioperative depression (n 37), although all threesubjects diagnosed with depression at 12 months had a juniorhigh school education or less.

DISCUSSION

This study is the first to prospectively assess depression andpsychological distress in NSCLC patients during a 1-year periodafter curative resection and to identify predictors of theirpsychological outcome at 12 months after surgery. Our studywas designed to obtain comprehensive data concerning depres-sion by using a rigorous diagnostic method and reliable, valid,self-administered instruments. However, most measurements atbaseline, including measures of depression and psychologicaldistress, could not be obtained before surgery for NSCLC,thereby limiting the comparison of overall degree of changefrom preoperative psychological functioning. Moreover, weshould indicate that recommendations for psychiatric consulta-tion to the attending physicians in 22 cases at 1 or 3 months aftersurgery and psychiatric interventions in six cases may haveaffected the prevalence of depression and psychological distress.The high attrition rate (19.1%) together with the fact that 17(41.5%) of the 41 surviving nonparticipants refused to partici-pate for reasons that included psychological burden mean thatthe prevalence of depression and psychological distress at 3months and 12 months may have been underestimated.

Despite several limitations, our finding showed that theprevalence of depression did not significantly change during theyear after surgery (range, 4.7% to 8.0%). This finding appears to besupported by the finding that the NSCLC patients psychologicaldistress measured by POMS-TMD did not change during the yearafter surgery, even after the six patients referred to the PsychiatryDivision or the 22 patients given recommendations for psychiatricconsultation were excluded. Even though the prevalence of depres-

Table 3. Predictors of Depression and Psychological Distress of NonSmall-Cell Lung Cancer Patients at 12 Months After Curative Resection:Logistic and Multiple Linear Regression Analysis

Dependent Variables/Independent Variables B SE OR 95% CI P

Depression (n 212)Depression at 1 month 2.08 0.83 7.98 1.58-40.39 .012Junior high school education or less 2.35 0.84 10.50 2.04-54.10 .005

R2 SEStandardized

Coefficient P

POMS-TMD (n 210) .272POMS-TMD at 1 month .241 0.06 0.47 .001History of depression between the diagnosis of

lung cancer and surgery.031 4.60 0.18 .003

Abbreviation: POMS-TMD, Profiles of Mood Disturbance-total mood disturbance.

Table 4. Predictors of Psychological Distress at 12 Months After CurativeResection of NonSmall-Cell Lung Cancer in Patients With and Without

Perioperative Depression: Multiple Linear Regression Analyses (n 210)

Dependent Variables/Independent Variables R2 SEStandardized

Coefficient P

Patients without perioperative depression(n 174)

POMS-TMD .251POMS-TMD at 1 month .251 0.07 0.50 .001

Patients with perioperative depression(n 36)

POMS-TMD .278Junior high school education or less .278 10.10 0.53 .001

NOTE. Perioperative depression means a depression episode between the diag-nosis of lung cancer and surgery or at 1 month after curative resection. Four patientshad a depression episode both between the diagnosis of lung cancer and surgeryand at 1 month.

Abbreviation: POMS-TMD, Profiles of Mood States-total mood disturbance.

74 UCHITOMI ET AL

sion was much lower than among lung cancer patients as a whole,including patients with advanced and small-cell lung cancer(15% to 44%),21-25 it was not low enough to be considerednegligible (0.9% to 3.7% in the general population).30-32

Although there have been no long-term studies of depressionafter surgery for NSCLC, one study5 examined global QOLaccording to the Quality of Life Index43 before and 1, 3, 6, and9 months after surgery in 117 consecutive subjects who under-went thoracotomy for a certain or presumptive diagnosis of lungcancer. Dales et al5 observed deterioration of QOL during thefirst 3 months postoperatively in those with a final diagnosis ofcancer (n 91) and found that although their QOL rebounded toits preoperative level, it did not reach the level of those in whomthe final diagnosis was not cancer (n 26). Our results aresimilar to their findings in that QOL did not reach the level ofthose without a final diagnosis of cancer during the 9 monthsafter surgery. Neither depression nor QOL of the NSCLCpatients after curative resection appears to decrease spontane-ously. Depression should be assessed repeatedly and should notbe underrecognized even after curative resection. Because de-pression at 1 month was a significant predictor, easy self-administered screening tools, such as the HADS,44,45 might bebeneficial to patients during the first year after successfulsurgical treatment of NSCLC.

Another noteworthy finding in this study was that the preva-lence of depression did not significantly decrease during the yearafter surgery, even though only three (18%) of the 17 patientswith depression at 1 month after surgery were diagnosed withdepression at 12 months. Seven (70%) of the 10 patients withdepression at 12 months were diagnosed for the first time. In twostudies that prospectively assessed depression before and aftertreatment for inoperable NSCLC,24, 25 the clinical course of thedepression could be explained in part by a transient reaction tothe diagnosis and treatment of cancer, by persistence of thereaction over time, or by worsening of the PS and the develop-ment of pain and dyspnea. In contrast, the results of our studyshowed a significant reduction of PS, pain, and dyspnea duringthe year after surgical treatment of early NSCLC; these findingsare consistent with previous reports.5-7 One possible explanationfor the finding that in most of the patients diagnosed withdepression at 12 months it was diagnosed for the first time, is thatjunior high school education or less was a significant predictor ofdepression, although it became a marginal and not significantpredictor after subjects with depression at 1 month were ex-cluded (P .053, n 195). The results of the study indicate thatmedical professionals should pay careful attention to less-educated NSCLC patients after curative resection.

There was no change in the NSCLC patients psychologicaldistress measured by POMS-TMD during the year after surgery,although all six subscale scores of the NSCLC patients in thisstudy were below the mean scores of healthy Japanese popula-tions 60 years of age or more on the Japanese version of thePOMS.41 Despite the significant reduction in PS to the preoper-ative level, both the anger-hostility and tension-anxiety score ofthe POMS increased during the year after surgery; however,there was a favorable change in the vigor-activity score. Becausea history of depression between the diagnosis of lung cancer andsurgery was a significant predictor, on the basis of the results ofour study as a whole, repetitive perioperative assessments ofdepression as well as careful attention to less-educated patients

might lead to early detection of and early treatment for depres-sion, resulting in amelioration of depression and psychologicaldistress during the year after curative resection of NSCLC.

Although it has been recommended that cancer patients beroutinely screened for distress,46 accumulating data indicate thatscreening programs for depression consume considerable re-sources and are not an efficient means of improving the mentalhealth outcome of medical patients.47 If the sensitivity (91.5%)and the specificity (65.4%) of the HADS used to screen fordepression in Japanese cancer patients45 were applied in thisstudy, the positive predictive value would be 25.3% for depres-sion between the diagnosis of lung cancer and surgery and 19.3%for depression at 1 month after surgery, and at least 2.5 and 2.4patients/mo, respectively, would have to be interviewed bymedical professionals with knowledge of the diagnostic proce-dures for depression to make routine screening efficient enoughto be practical.

Being informed of a diagnosis of lung cancer could beconsidered to be an acute stressful life event, as would adiagnosis of any serious physical illness. According to theseverity of psychosocial stressors scale of the DSM-III-R,26

being diagnosed with a serious illness is graded as an extremelysevere psychosocial stressor, the same as the death of a spouse;the grade is between a severe event (eg, divorce) and acatastrophic event (eg, death of a child). Previous studies on thedeath of a spouse have demonstrated a high prevalence of majordepression that decreased over time but remained higher than inthe controls (33% to 35% at 1 month after the loss, 23% to 25%at 2 months, 16% to 17% at 13 months, respectively).48,49 Theprevalence of major depression in this study appears to be muchlower than following an extremely severe psychosocial stressorsuch as the death of a spouse, and it shows a tendency todecrease during the year after surgery.

There were several limitations to this study. First, there wassampling bias, because the results were obtained from only oneinstitution, which was a teaching cancer center hospital. Second,recommendations for psychiatric consultation were made to theattending physicians. Moreover, there was a high attritionratemany nonparticipants refused to participate for reasonsthat included psychological burden, and psychiatric interventionoccurred for six patients with depression. In fact, five of the sixpatients referred to the Psychiatry Division were not diagnosedwith depression at 12 months after surgery. Therefore, theprevalence of depression and psychological distress may havebeen underestimated as a result. Thus, if the depression in thesefive patients had persisted without psychiatric interventions, theprevalence of depression, major depression, and minor depres-sion at 3 and 12 months would have been as high as 6.1%, 3.8%,and 2.4%, respectively, at 3 months, and 7.1%, 2.8%, and 4.2%,respectively, at 12 months. Third, it was disappointing that mostmeasurements at baseline, including measures of depression andpsychological distress, could not be obtained before surgery forNSCLC, thereby limiting the comparison of overall degree ofchange from preoperative psychological functioning. The shortpreoperative period results from a difference in our institutionsmedical procedure. Preoperative psychological status in thisstudy may not really reflect baseline mental health because of thehigh stress of the moment. Fourth, the prevalence of majordepression in Asian countries is generally lower than in westerncountries, possibly because of cross-cultural differences (ie,

75DEPRESSION AFTER LUNG CANCER SURGERY

social stigma, cultural reluctance to endorse mental symptoms,and low divorce rate).50-52 The interpretation of the results in thisstudy is cautious. Fifth, depression may have been overestimatedbecause we elected to use an inclusive diagnostic approach. Thisapproach includes somatic symptoms, regardless of whether therater judges that the symptom is caused by medical or psycho-logical causes, prevents underdiagnosis of depression, and isreliable because of the high interrater agreement.53 Becauseother approaches do not offer a clear significant advantage inmeasuring depression and the need for treatment,54 the inclusiveapproach may be recommended with limitations in the clinicaloncology setting. Finally, although the simple four-point verbalpain rating scale is the most widely used in the clinical context,the fact that we did not use a pain rating scale with highersensitivity to change may have adversely affected the pain anddyspnea assessments in this study.55

Depression is not routinely assessed even in patients withunresectable NSCLC, although the majority of the previousextensive research on QOL has addressed the need for psycho-social support for such patients. Furthermore, patients faceuncertainty and fear of recurrence after curative resection, even

though surgical treatment for early-stage NSCLC is generallyconsidered curative. This study revealed that the prevalence ofdepression did not reach a negligible level and did not changeafter curative resection for NSCLC. It also provided informationindicating that perioperative depression and less-educated statuswere significant predictors of depression at 12 months aftersurgery. Overall, the study indicates that the psychological statusof resectable NSCLC patients needs to be systematically ad-dressed during their overall rehabilitation, including pain anddyspnea management. In addition, this study identified an areathat is under studied in the literature regarding cancer survivor-ship, especially in the context of curative treatment. In the future,a randomized trial of coordinated psychosocial interventions onthe basis of patient screening and treatment should be performedwith the aim of ameliorating depression and psychologicaldistress during the year after curative resection.

ACKNOWLEDGMENT

We are grateful to the patients and the physicians of the ThoracicOncology Division, National Cancer Center Hospital East, and for aGrant-in-Aid for Cancer Research from the Japanese Ministry of Health andWelfare.

REFERENCES

1. Parkin DM, Pisani P, Ferlay J: Estimates of the worldwide incidence of25 major cancers in 1990. Int J Cancer 80:827-841, 1999

2. Pisani P, Parkin DM, Bray F, et al: Estimates of the worldwidemortality from 25 cancers in 1990. Int J Cancer 83:18-29, 1999

3. Bernhard J, Ganz PA: Psychosocial issues in lung cancer patients (part1). Chest 99:216-223, 1991

4. Montazeri A, Gillis CR, McEwen J: Quality of life in patients withlung cancer: A review of literature from 1970 to 1995. Chest 113:467-481, 1998

5. Dales RE, Belanger R, Shamji FM, et al: Quality-of-life followingthoracotomy for lung cancer. J Clin Epidemiol 47:1443-1449, 1994

6. Nou E, Aberg T: Quality of survival in patients with surgically treatedbronchial carcinoma. Thorax 35:255-263, 1980

7. Hamelmann H, Thermann M, Muller-Schwefe T, et al: Surgicallytreated bronchial carcinoma patients: Results of systematic follow-up.Thorac Cardiovasc Surg 31:41-44, 1983

8. Naruke T, Tsuchiya R, Kondo H, et al: Prognosis and survival afterresection bronchogenic carcinoma based on the 1997 TNM-staging classifi-cation: The Japanese experience. Ann Thorac Surg 71:1759-1764, 2001

9. Cella DF: Quality of life outcomes: Measurement and validation.Oncology 11:S233-S246, 1996 (suppl)

10. Vissor MRM, Smets EMA: Fatigue, depression and quality of life incancer patients: How are they related? Support Care Cancer 6:101-108, 1998

11. Harrison J, Haddad P, Maguire P: The impact of cancer on keyrelatives: A comparison of relative and patient concerns. Eur J Cancer31A:1736-1740, 1995

12. Colleoni M, Mandala M, Peruzzotti G, et al: Depression and degree ofacceptance of adjuvant cytotoxic drugs. Lancet 356:1326-1327, 2000

13. Buccheri G: Depressive reactions to lung cancer are common andoften followed by a poor outcome. Eur Respir J 11:173-178, 1998

14. Faller H, Bulzebruck H, Drings P, et al: Coping, distress, and survivalamong patients with lung cancer. Arch Gen Psychiatry 56:756-762, 1999

15. Maguire P, Booth K, Elliott C, et al: Helping health professionalsinvolved in cancer care acquire key interviewing skills: The impact ofworkshops. Eur J Cancer 32A:1486-1489, 1996

16. Razavi D, Delvaux N, Hopwood P: Improving communication withcancer patients. A challenge for physicians. Ann N Y Acad Sci 809:350-360,1997

17. Passik SD, Dugan W, McDonald MV, et al: Oncologists recognitionof depression in their patients with cancer. J Clin Oncol 16:1594-1600, 1998

18. Derogatis LR, Morrow GR, Fetting J, et al: The prevalence ofpsychiatric disorders among cancer patients. JAMA 249:751-757, 1983

19. Massie MJ, Holland JC: Depression and the cancer patient. J ClinPsychiatry 51:12-17, 1990

20. Minagawa H, Uchitomi Y, Yamawaki S, et al: Psychiatric morbidityin terminally ill cancer patients. Cancer 78:1131-1137, 1996

21. Hughes JE: Depressive illness and lung cancer: II. Follow-up ofinoperable patients. Eur J Surg Cancer 11:21-24, 1985

22. Ginsburg ML, Quirt C, Ginsburg AD, et al: Psychiatric illness andpsychosocial concerns of patients with newly diagnosed lung cancer. CanMed Assoc J 152:701-708, 1995

23. Montazeri A, Milroy R, Hole D, et al: Anxiety and depression inpatients with lung cancer before and after diagnosis: Findings from apopulation in Glasgow, Scotland. J Epidemiol Community Health 52:203-204, 1998

24. Hopwood P, Stephens RJ: Depression in patients with lung cancer:Prevalence and risk factors derived from quality-of-life data. J Clin Oncol18:893-903, 2000

25. Akechi T, Okamura H, Nishiwaki Y, et al: Psychiatric disorders andassociated and predictive factors in patients with unresectable nonsmall celllung carcinoma. Cancer 92:2609-2622, 2001

26. American Psychiatric Association. Diagnostic and Statistical Manualof Mental Disorders (ed 3), Revised. Washington, DC, American PsychiatricPress, 1987

27. Zigmond AS, Snaith RP: The hospital anxiety and depression scale.Acta Psychiatr Scand 67:361-370, 1983

28. Spitzer RL, Williams JBW, Gibbon M, et al: Structured ClinicalInterview for DSM-III-R. Washington, DC, American Psychiatric Press,1990

29. Uchitomi Y, Mikami I, Kugaya A, et al: Depression after successfultreatment for nonsmall cell lung carcinoma. Cancer 89:1172-1179, 2000

30. Hasegawa K: The epidemiological study of depression in late life.J Affect Disord S3-S6, 1985 (suppl 1)

31. Regier DA, Boyd JH, Burke JD Jr, et al: One-month prevalence ofmental disorders in the United States. Based on five Epidemiologic Catch-ment Area sites. Arch Gen Psychiatry 45:977-986, 1988

32. Steffens DC, Skoog I, Norton MC, et al: Prevalence of depression andits treatment in an elderly population: The Cache County study. Arch GenPsychiatry 57:601-607, 2000

33. Kaasa S, Malt U, Hagen S, et al: Psychological distress in cancerpatients with advanced disease. Radiother Oncol 27:193-197, 1993

34. Cella DF, Orofiamma B, Holland JC, et al: The relationship ofpsychological distress, extent of disease, and performance status in patientswith lung cancer. Cancer 60:1661-1667, 1987

76 UCHITOMI ET AL

35. Dugan W, McDonald MV, Passik SD, et al: Use of the ZungSelf-Rating Depression Scale in cancer patients: Feasibility as a screeningtool. Psychooncology 7:483-493, 1998

36. Wenzel LB, Fairclough DL, Brady MJ, et al: Age-related differencesin the quality of life of breast carcinoma patients after treatment. Cancer86:1768-1774, 1999

37. Revenson TA, Wollman CA, Felton BJ: Social supports as stressbuffers for adult cancer patients. Psychosom Med 45:321-331, 1983

38. Mountain CF: A new international staging system for lung cancer.Chest 89:225S-233S, 1986 (suppl)

39. McNair D, Lorr M, Droppleman L: EITS Manual for the Profile ofMood States. San Diego, CA, Educational and Industrial Testing Service,1971

40. Kathol RG, Mutgi A, Williams J, et al: Diagnosis of major depressionin cancer patients according to four sets of criteria. Am J Psychiatry147:1021-1024, 1990

41. Yokoyama K, Araki S, Kawakami N, et al: Production of the Japaneseedition of Profile of Mood States (POMS): Assessment of reliability andvalidity. Jpn J Public Health 37:913-918, 1990

42. Hariharan M, VanNoord T, Greden JF: A high-performance liquid-chromatographic method for routine simultaneous determination of nicotineand cotinine in plasma. Clin Chem 34:724-729, 1988

43. Spitzer WO, Dobson AJ, Hall J, et al: Measuring the quality of life ofcancer patients: A concise QL-index for use by physicians. J Chronic Dis34:585-597, 1981

44. Hopwood P, Howell A, Maguire P: Screening for psychiatric mor-bidity in patients with advanced breast cancer: Validation of two self-reportquestionnaires. Br J Cancer 64:353-356, 1991

45. Kugaya A, Akechi T, Okuyama T, et al: Screening for psychologicaldistress in Japanese cancer patients. Jpn J Clin Oncol 28:333-338, 1998

46. Holland JC: Preliminary guidelines for the treatment of distress.Oncology 11:109-114, 1997

47. Gilbody SM, House AO, Sheldon TA: Routinely administered ques-tionnaires for depression and anxiety: Systematic review. BMJ 322:406-409,2001

48. Bornstein PE, Clayton PJ, Halikas JA, et al: The depression ofwidowhood after thirteen months. Br J Psychiatry 122:561-566, 1973

49. Zisook S, Shuchter SR: Depression through the first year after thedeath of a spouse. Am J Psychiatry 148:1346-1352, 1991

50. Weissman MM, Bland RC, Canino GJ, et al: Cross-national epide-miology of major depression and bipolar disorder. JAMA 276:293-299, 1996

51. Compton WM, Helzer JE, Hwu HG, et al: New methods in cross-cultural psychiatry: Psychiatric illness in Taiwan and the United States. Am JPsychiatry 148:1697-1704, 1991

52. Nakane Y, Ohta Y, Radford M, et al: Comparative study of affectivedisorders in three Asian countries: II. Differences in prevalence rates andsymptom presentation. Acta Psychiatr Scand 84:313-319, 1991

53. Koenig HG, Pappas P, Holsinger T, et al: Assessing diagnosticapproaches to depression in medically ill older adults: How reliably canmental health professionals make judgments about the cause of symptoms?J Am Geriatr Soc 43:472-478, 1995

54. Koenig HG, George LK, Peterson BL, et al: Depression in medicallyill hospitalized older adults: Prevalence, characteristics, and course ofsymptoms according to six diagnostic schemes. Am J Psychiatry 154:1376-1383, 1997

55. Caraceni A, Cherney N, Fainsinger R, et al: Pain measurement toolsand methods in clinical research in palliative care: Recommendations of anexpert working group of the European Association of Palliative Care. J PainSymptom Manage 23:239-255, 2002

77DEPRESSION AFTER LUNG CANCER SURGERY